Network graph outlier detection for identifying suspicious behavior

Abstract

A computer-implemented method for detecting suspicious or fraudulent insurance claim filings may include receiving a list of individuals who file insurance claims; receiving a list of contacts for each individual; receiving information regarding relationships between the contacts; forming a plurality of ego networks that each include a central hub, a plurality of nodes, and a plurality of edges; determining a number of nodes for each ego network; determining a number edges for each ego network; forming a plurality of data points from the numbers of nodes and the numbers of edges; and calculating a distance of each data point from a predetermined normal relationship function to facilitate identifying outliers that warrant investigation or may be associated with insurance claim buildup.

Having thus described various embodiments of the invention, what is claimed as new and desired to be protected by Letters Patent includes the following: 
     
         1 . A computer-implemented method for detecting outliers, the method comprising the following steps, wherein each step is performed by a processor of a computing device:
 receiving, from a memory element, a list of individuals who file insurance claims;   receiving, from the memory element, a list of contacts for each individual;   receiving, from the memory element, information listing relationships between two or more of the contacts and between each contact and the individual;   forming a plurality of ego networks, one ego network formed for each individual with each ego network including a central hub representing the individual, a plurality of nodes with each node representing a contact, and a plurality of edges with each edge representing a relationship between one contact and the individual or between two contacts;   determining a number of nodes for each ego network;   determining a number of edges for each ego network;   forming a plurality of two-dimensional data points from the numbers of nodes and the numbers of edges, with each data point representing one ego network and the number of nodes of the ego network forming an x-coordinate of the data point and the number of edges of the ego network forming a y-coordinate of the data point;   developing, using a computer learning system, a mathematical function defining a normal relationship between edges and nodes for each ego network, wherein developing the mathematical function comprises applying curve fitting to the data points;   determining a distance of each data point from the mathematical function defining a normal relationship between edges and nodes for each ego network to facilitate identifying the outliers; and   displaying, on a display device, names or ID numbers of the outliers.   
     
     
         2 . The computer-implemented method of  claim 1 , further comprising determining whether the distance of each data point is greater than a threshold. 
     
     
         3 . The computer-implemented method of  claim 2 , further comprising reporting the individuals associated with the data points whose distance is greater than the threshold. 
     
     
         4 . The computer-implemented method of  claim 1 , wherein determining the number of nodes for each ego network includes counting each contact as one node. 
     
     
         5 . The computer-implemented method of  claim 1 , wherein determining the number of edges for each ego network includes counting each relationship between the individual and one contact as one edge and each relationship between two contacts as one edge. 
     
     
         6 . (canceled) 
     
     
         7 . The computer-implemented method of  claim 1 , wherein the mathematical function defining the normal relationship is a linear function. 
     
     
         8 . A computer-implemented method for detecting outliers, the method comprising the following steps, wherein each step is performed by a processor of a computing device:
 receiving, from a memory element, a list of individuals who file insurance claims;   receiving, from the memory element, a list of contacts for each individual;   receiving, from the memory element, information listing relationships between two or more of the contacts and between each contact and the individual;   forming a plurality of ego networks, one ego network formed for each individual with each ego network including a central hub representing the individual, a plurality of nodes with each node representing a contact, and a plurality of edges with each edge representing a relationship between one contact and the individual or between two contacts;   determining a number of nodes for each ego network;   determining a number of edges for each ego network;   forming a plurality of two-dimensional data points from the numbers of nodes and the numbers of edges, with each data point representing one ego network and the number of nodes of the ego network forming an x-coordinate of the data point and the number of edges of the ego network forming a y-coordinate of the data point; and   developing, using a computer learning system, a linear mathematical function defining a normal relationship between edges and nodes for each ego network, wherein developing the linear mathematical function comprises applying linear regression to the data points;   determining a distance of each data point from the linear mathematical function defining a normal relationship between edges and nodes for each ego network;   determining whether the distance of each data point is greater than a threshold; and   reporting the individuals associated with the data points whose distance is greater than the threshold to facilitate identifying the outliers, wherein reporting the individuals associated with the data points whose distance is greater than the threshold comprises displaying, on a display device, names or ID numbers associated with the individuals associated with the data points whose distance is greater than the threshold.   
     
     
         9 . The computer-implemented method of  claim 8 , wherein determining the number of nodes for each ego network includes counting each contact as one node. 
     
     
         10 . The computer-implemented method of  claim 8 , wherein determining the number of edges for each ego network includes counting each relationship between the individual and one contact as one edge and each relationship between two contacts as one edge. 
     
     
         11 . (canceled) 
     
     
         12 . A computer-implemented method for detecting outliers, the method comprising the following steps, wherein each step is performed by a processor of a computing device:
 determining and/or receiving a list of individuals;   determining and/or receiving a list of contacts for each individual;   determining and/or receiving information listing relationships between the contacts and/or the individuals;   forming or generating a plurality of ego networks, one ego network formed for each individual with each ego network including a central hub representing the individual, a plurality of nodes with each node representing a contact, and a plurality of edges with each edge representing a relationship between one contact and the individual or between two contacts;   determining a number of nodes for each ego network;   determining a number of edges for each ego network;   forming a plurality of two-dimensional data points from the numbers of nodes and the numbers of edges, with each data point representing one ego network and the number of nodes of the ego network forming an x-coordinate of the data point and the number of edges of the ego network forming a y-coordinate of the data point;   developing, using a computer learning system, a mathematical function defining a normal relationship between edges and nodes for each ego network, wherein developing the mathematical function comprises applying curve fitting to the data points;   determining a distance of each data point from the the mathematical function defining a normal relationship between edges and nodes for each ego network to facilitate identifying the outliers; and   displaying, on a display device, names or ID numbers of the outliers.   
     
     
         13 . The method of  claim 12 , wherein the individuals are associated with insurance products or services or financial products or services. 
     
     
         14 . The method of  claim 12 , wherein the individuals are associated with filing insurance claims and an abnormal distance calculated for a data point is indicative of insurance claim fraud. 
     
     
         15 . The method of  claim 12 , wherein the individuals are medical services providers that submit insurance claims on behalf of patients. 
     
     
         16 . The method of  claim 12 , wherein the individuals are construction workers or companies that repair damaged insured homes.